Overview
According to data from the World Health Organization, approximately 160 million people worldwide are edentulous. The incidence increases with age, and the proportion of edentulous patients is higher in the population aged 60 and above. Loss of teeth or edentulism can affect facial appearance, causing people to feel self-conscious and loss confidence in social situations, and even lead to psychological illnesses. Therefore, edentulous patients not only pay close attention to the recovery of oral function but also attach great importance to facial contour improvement. For a long time, due to technological limitations, clinicians have been unable to depict the changes in facial contour after implant placement for patients before surgery. However, with the development of artificial intelligence technology, deep learning-based methods for predicting soft tissue facial deformation have made this mission a possibility. This study established a multi-modal dataset for edentulous patients before and after implant restoration to lay the foundation for predicting facial contour changes after implant treatment. A graph generative adversarial network based on multi-modal data was proposed to achieve fast and high-precision facial contour prediction. To address the common challenges of slow computation and excessive computational resource consumption in current triangular mesh deformation simulation methods, this project innovatively proposed a graph generative adversarial network that uses multi-modal data and incorporates self-attention mechanisms to achieve fast and high-precision facial contour prediction for edentulous patients after implant restoration.
Eligibility
Inclusion Criteria:
- Patients with complete edentulism,
- aged 50 years or above,
- in good physical health,
Exclusion Criteria:
- patients who refuse to participate in the study，
- patients who cannot undergo facial scanning.